In a data center, a physical network is established by connecting a plurality of physical servers in a wired manner, and virtual machines or virtual switches are operated using the physical servers included in the physical network. With such a configuration, virtual networks have been increasingly established. Even in a system in which the virtual network is established in the physical network, there are some cases where a plurality of paths is used in communication between devices in order to secure a communication speed or measures against a system failure.
When communication is performed between the virtual machines included in the virtual network, communication is performed between the physical servers in which the virtual machines that perform communication operate, respectively. Thus, both a case where causes of the failure are present in the virtual network and a case where causes of the failure are present in the physical network are considered in the system in which the virtual network is established in the physical network. Accordingly, when the measures against the system failure are set up, an operator is to specify whether the causes of the failure are attributed to the failure in the physical network or are attributed to the problem in the virtual network. Here, the operator specifies whether or not the failure occurs in the physical network, and determines that a communication failure occurs due to the failure in the virtual network in a case where the failure does not occur in the physical network.
However, there are some cases where it is difficult to find a location in which the failure occurs by simply performing real-time monitoring or log analysis in the physical server included in the physical network. Thus, there are some cases where whether or not packet loss occurs is determined using a check packet.
As related arts, a test method using a packet acquired by adding path information to a header portion to which destination information is set has been suggested (for example, Japanese Laid-open Patent Publication No. 9-270822). In this test method, a relay node transmits a packet transmitted from a transmission source based on routing information, and a device as a transmission source determines a state of a communication path based on whether or not a response packet is received from a destination specified using the destination information. A system in which a node that receives a frame transmitted from a management computer transmits a received frame based on the routing information indicating a delivery path of the frame in a communication network and transmits a response to the management computer also has been suggested. In this system, the management computer specifies a location in which the failure occurs based on a reception state of the response from the node included in the network (for example, International Publication Pamphlet No. WO 2010064532). A communication method of enabling restoration even though the packet loss occurs in any one path by generating a redundant packet if the packet is received and transmitting the received packet and the redundant packet over different paths to a network as a destination has been suggested (for example, Japanese Laid-open Patent Publication No. 2004-274703).